-1

I have to make a project college. I have to come with some application of signals that I can implement in MATLAB.

Is it possible to separate two voices from an audio signal, say a .waw file ?

I mean, using techniques from a first course in digital signal processing (DFT, Spectogram, Cepstrum).

Esteban Filardi
  • 232
  • 2
  • 7

1 Answers1

0

(i dont have enough points to post comments like everyone else)

The problem you are referring to is commonly called "the cocktail problem algorithm" here is a little link I found https://stackoverflow.com/questions/20414667/cocktail-party-algorithm-svd-implementation-in-one-line-of-code

I actually discovered this algorithm, in that same online course on coursera a few weeks ago. This article also has some links to research.

If you want to see it (I recommend you do) just sign up at coursera for the stanford machine learning class. The video is the 4th entry from week 1 introduction at approximately 5mins 30 seconds in, it's titled "unsupervised learning". I don't think I can legally post a link to the video, but since it's free, why not take a look. At this point it's important to say other than those few minutes at the end of the video clip, I don't think he discusses the cocktail party algorithm again. So there is no need to scour through hours of videos and lectures, you won't find anything (at least in the currently posted content, I haven't finished the course yet)

andrew
  • 730
  • 3
  • 9
  • 1
    hey, I originally referenced the wrong video, it isn't the 3rd its the 4th update in bold – andrew Feb 18 '15 at 19:24